Need Help Calibrating a Precision Level

D

ddickey

H-M Supporter - Gold Member
H-M Supporter Gold Member
Joined
Apr 21, 2016
Messages
3,570
This is the process I'm using. Let me know what I'm doing wrong here.
I find a level spot on my surface plate. Mark the spot and rotate 180°.
Level does not repeat, this tells me it is not calibrated.
Okay so now I'm trying to find a spot on the plate where the bubble is exactly opposite when you rotate the level 180°. This will tell you that you've found a true level spot. Once you find this you can then use the adjusting screw to center the bubble.
The problem I'm having is I can not find a spot anywhere where the bubble repeats it's position exactly on the opposite side when you rotate the level 180°.
My post is probably hard to follow. Please ask questions and I'll try to clarify.
 
If the bubble reads the same in both directions then it is calibrated. Ideally you would have an adjustable flat surface that could be adjusted for tilt so that the level would read level when flipped. So when leveling my lathe, I always double check the accuracy of my level by flipping it and taking a reading, if the lathe is offe by one division it will read the same when the level is flipped 180 degrees (i.e. off one division in the same direction).
 
And if you flip while the bubble is in the center it will stay in the center if calibrated. Mine does not probably about three divisions off. I'm very careful to flip exactly the same spot.
 
It does not matter if your surface is level. When you rotate the level 180 on exactly the same spot you need to adjust the level until you do get exactly the same reading at 0 and 180, then the level is calibrated. That same reading does not have to be centered, just the same both ways. Once the level is calibrated then you can use it to level the surface plate if you so desire. A surface plate does not have to be level to do its job unless you are taking readings using a level on a part that is sitting on the plate.

When using a level I always make sure that I run the bubble from one end to the other before taking readings to insure that there is the same file of fluid on all parts of the glass. The bubble on oil will have a different amount of stickines than the bubble on bare glass.
 
The surface you are using does not have to be level IF the amount that it is off is within the range of the level. For example, if your level has 3 lines to either side and the bubble is in this range then you can work with that but if it is more than that then is the level off or is the surface off? The reason this matters is because it can be really confusing as to what is causing the level to not reproduce its reading when you flip it - is it the level itself or the surface you're using?

I am attaching a procedure that should help you to clear all of this up. I use an adjustable melamine-coated table to calibrate my precision level, not a surface plate. The table is elevated on three feet so I can clamp a small aluminum flat square to the table anywhere I choose. This square serves as a register so the level can be flipped 180 degrees and repositioned precisely in the same spot. I use a round bullseye-type level to get the table close to level and begin the calibration procedure. For me, this table really simplifies what is otherwise an often long and frustrating process.

My level is a Kinex 11" level that is accurate to 0.02mm/M so it is really sensitive. It can take 30 seconds or more for the bubble to stop moving. The table lets me verify calibration quickly and I would not do without it.
 

Attachments

  • Engineers-Level-Calibration-Instructions.pdf
    38.8 KB · Views: 24
Last edited:
Thanks Mike,
I lost that when I switched computers a while back.
I watched this video and he complicates matters by trying to find an exact error on both sides when rotated 180°. This was making it really difficult as mine is very sensitive.
Currently mine is three divisions to the left and flipped it is 1 division to the left. That tells me surface is not level and level is out of cal by one division. If I adjust so that both readings are two to the left then I should be good.

EDIT: Forgot to add the video link.
 
Last edited:
I had a student that was top of his class at MIT show me this way. (it is possible to teach an old dog new tricks) Your surface plate does not need be level. You set your stoned level on a clean plate in any place and slowly rotate the level in a circle or spinning it 360 degree's on that same spot and eventually you will find a true level spot. Then lay a straight-edge rule next to it with one end of the rule lined up with the end of level. If you think about it the plate could be at a 45deg angle, but if you spin the level 90 degs it will read level.....once you have the end of rule lined up with the one end flip the level 180 degree's and it should repeat to the same place. if not adjust the bubble 1/2 the distance it is off, and flip it again and check. I only adjust on one side of the flip. So flip, adjust and flip back, read, flip and adjust again.

Another way would be to slide in a feeler gage under one end until it reads level and do the flip / adjust test. Many times I scrape the bottom of Starrett 198 and 199 levels as they will pivot in the middle because most people seat the level on a dirty surface and it wears the ends low. You need a grade A or better plate to test the pivot and it is easy to check. Once you have the level with-in 1 line which is .0005" on a 199, you push down on the ends and watch the bubble and it should not move. if it does it needs to be scraped and when I scrape them I usually leave the middle 40% low .0005" unless I am using the level on narrow surfaces and then I scrape the perfectly flat.

One can level the plate first as the others say too. But if your in a hurry you can use an unlevel plate.
 
Last edited:
ddickey said:
Thanks Mike,
I lost that when I switched computers a while back.
I watched this video and he complicates matters by trying to find an exact error on both sides when rotated 180°. This was making it really difficult as mine is very sensitive.
Currently mine is three divisions to the left and flipped it is 1 division to the left. That tells me surface is not level and level is out of cal by one division. If I adjust so that both readings are two to the left then I should be good.
Click to expand...

ddickey, next week I could drive down and show you how, can bring a scraper and one of my 199's. I am super busy now, and seeing we are with-in 30 miles or so. :)
 
You must log in or register to reply here.
Back
Top